SPANISH NEOGENE RHINOCEROSES /n’ESPERANZA CERDENO Abstract. Nine genera and fourteen species are recognized among the rhinocerotid remains of the Spanish Miocene and Lower Pliocene. Protaceratherium mumtiim and Prosautorhinus sp. are reported in Spain for the first lime, and the species AUcomops alfanihrense has been published recently as a final form in the A. simorrense evolutionary lineage. The overall diversity of Rhinocerotidae in the Neogene record of Spain is generally much greater than that of the Ec]uidae. Palaeoclimatic curves for the Spanish Neogene show how rhinoceros diversity depended on the relative humidity and the temperature. This paper reviews the rhinocerotids of the Spanish Neogene, with data from the author’s Ph.D. (Cerdeno 1989). Up to now, among Spanish Neogene rhinoceroses, only the fossil remains from the Valles-Penedes basin had been studied in detail (Santafe 1978c/), and they alone have mainly been used as a point of reference. The fossil material of rhinoceroses comes from 45 Spanish sites, with more than 2000 identified bones, but including only a few cranial remains. These sites have a wide geographical and temporal distribution. They are located in several sedimentary basins; Ebro, Duero, Calatayud-Teruel, with two main areas (Daroca and Teruel), Tajo, and the Eastern and Betic basins. The chronological distribution ofthe localities ranges through the Miocene and Lower Pliocene (Ruscinian). As can be seen in Text-figure I, the fossil record is continuous through that time, and only the base ofthe middle Aragonian is unknown, which is the case for all macromammal groups, not just rhinocerotids. The biozones I have used are those of Mein (1975) and Daams and Freudenthal (1981). Anatomical abbreviations are; E incisor; Me, metacarpal; P, premolar. SYSTEMATIC PALAEONTOLOGY Family rhinocerotidae Owen, 1845 Rhinocerotidae indet. I The presence ofa rhinocerotid in the Lower Miocene ofCctina dc Aragon, Zaragoza, is documented only by a distorted Mclil whose size and proportions are close to those of the McIII from Paulhiac and Laugnac, France, that Bonis (1973) identified as Diceratlieriiim pleuroceros. This name corresponds to Pleiiroceros plewocerosas used by Antuncsand Ginsburg(1983), Protheroetal. (1986) and Guerin (1989). At themoment, it is not possible to assert that the bone from Cetina corresponds to that species, and it is advisable to classify it as Rhinocerotidae indet. I. Subfamily menoceratinae Prothero, Manning and Hanson, 1986 Protaceratherium minutwn (Cuvier, 1822-1825) The genus Protaceratherium has recently been included in the tribe Menoceratini (Heissig 1989) within the subfamily Aceratheriinae, but Prothero and Schoch (1989) consider it as a member of the subfamily Menoceratinae. Protaceratherium mimitum is the oldest rhinocerotid recorded in the Spanish Neogene. It is known from only two Lower MiocenesitesoftheCuenca province, Valquemado and Lorancadel Campo. The lormer corresponds to the MN 2h or Y zone in age and the latter to the MN 3r/ or Z (Ginsburg et al. 19876). Thespeciesisvery abundant, withmorethan400remainsin Loranca. Amongthem, thereisacranial fragment IPalaeontology, Vol. 35, Part 2, 1992, pp, 297-308. © The Palaeontologieal Assoeiation PALAEONTOLOGY, VOLUME 298 35 TABLE I. Biostratigraphical distribution of the European localities with Protaceratheriwn minutum, after Ginsburg et al. (1987), Antunes and Ginsburg (1983), Bonis (1973). 1, Mein (1975); 2, Daams and Meulen (1984). (1) (2) Spain France Others 4a Artenay 36 A Lisboa I Beilleaux Chitenay 3u Z Loranca 26 Laugnac La Chaux Y Valquemado Selles/Cher Ulm la Moissac Budenheim 1 X Paulhiac Pechbonieu which coincides basically with the cranium from Budenheim, Germany, figured by Roman (1924), our specimen being wider in the middle of the occipital face. The small and slender P. minutum shows the upper premolars with a strong lingual cingulum, and there is a bridge of union between protoloph and metaloph whose development is variable. This variation can be observed in the sample from Loranca del Campo, as well as in the French remains from Selles-sur-Cher. Both cases also show the dificrent development ofthe crochet and crista ofthe upper premolars. Lower teeth from Valquemado show labial and lingual cingula only on P.,, while at Loranca cingula are present on all the premolars and some molars. The differences among the teeth and the postcranial skeleton of the studied samples ofP. minutum are mainly referred to size. Thisis greater in Loranca, where a slightly greater thickness can also be observed. When the Spanish material iscompared with the European data on P. mimitunu theclosest identity is found with theSelles-sur-Chersample. The French molarsarerelatively smallerthan thepremolarsand thesamealso occurs in the Beilleaux sample. The size ofthe Selles-sur-Cher specimens is closer to those from Valquemado than those from Loranca. This is expected since Selles-sur-Cher is supposed to be almost equivalent in age to the former and older than the latter (Table 1). The trend observed between the two Spanish sites (size increase and slenderness decrease through time) cannot be generalized because younger sites like Beilleaux, France, or Lisbon, Portugal, have smaller-sized specimens than Loranca, The maxillary fragment from Lisbon (Roman and Torres 1907) has been considered as a distinct species, Protaceratherium tagicunu because ofits smaller size, shorter upper PL and the disappearance of the lingual cingulum at the protocone level (Antunes and Ginsburg 1983). As already stated (Cerdeno 1989), these characters show great variation in the different studied samples. So, the size of the Lisbon teeth exceeds those from Beilleaux. On the other hand, the real length ofthe P^ cannot be measured because the specimen lacks theectoloph. The discontinuity ofthe lingual cingulum at the protocone as well as at the hypocone level has been observed in several specimens from Loranca, Selles-sur-Cher, Beilleaux, and Faluns (Touraine, Anjou). This all suggests that the remains from Lisbon do not constitute a distinct species, at least with the present data, and so they are best included in P. minutum. Concerning thepostcranial skeleton, the similarityis retained between Valquemado and Selles-sur-Cher and there is also a clear resemblance with the German remains from Budenheim (Roman 1924), whose dimensions are also smaller than in Loranca. In Western Europe, Protaceratherium minutum ranges from the earliest Lower Miocene (Pechbonieu and Cintcgabelle, France; Antunesand GinsburMgN1983) to theearly middle Aragonian (Artenay, France) (Table 1). Its best representation corresponds to the 2-3 or Y-Z-A biozones, with a variation in size that possibly reflects different habitats among the distinct areas where the species has been found. Later in the middle Aragonian another genus, Plesiacei\itherium, whose generic characteristics (Yan and Heissig 1986) are basically those of Protaceratherium, has been reported in Spain as well as in other Western European countries, 1 think both genera must be synonymous because it is just the larger size of Plesiaceralherium that marks the difference between them. So, Protaceratherium has priority over P/e.siuceratherium. Plesiaceratherium platyocion has been found in several middle Aragonian sites of Western CERDENO: SPANISH RHINOCEROSES 299 Europe, such as the Spanish sites of Buhol, Valencia (Belinchon 1987), and Can Mas, Barcelona (Santafe 1978«). It might be thought that this species is a descendant of Prolaceratlieriiini miniitum. Subfamily aceratheriinae Dollo, 1885 Aliconiops simoiTcnse (Lartet, 1851) Alicoriiops alfamhreuse Cerdeno and Alcala, 1989 Yan and Heissig (1986) and, more recently, Guerin (1989) and Heissig (1989) have already considered the species A. simon-ense as thegenus Aliconiops. Thedehnition ofthis genus would correspond basically with the subgeneric diagnosis (Ginsburg and Guerin 1979) which has been partly reviewed (Cerdeno 1989). One ofthe characteristics established by Ginsburg and Guerin (1979), the lack of MeV, can be clearly refuted by the existence of several well-developed MeV in the Spanish sites ofToril-3, Arroyo del Val, and Los Valles de Fuentiduena. At the moment, what really characterizes Aliconiops is the shortening ofthe limbs. This genus includes two species, A. simoircnse and a second one, A. alfanihren.se. recently described from the upper Vallesian of Spain (Cerdeno and Alcala 1989). A. siniorrense is well known from the upper Aragonian and Vallesian in Western Europe, but it is in Spain that it is best represented with more than 800 remains, among which are three cranial fragments, almost the only ones known for this species. It is widely distributed in the Valles-Penedes (Santafe I978«, 1978b), Duero, Tajo, and Calatayud-Daroca basins. Most ofthe sites are upper Aragonian and Los Valles de Fuentiduena, Nombrevilla, and Relea are lower Vallesian. Only in the Valles-Penedes basin has .4. siniorren.se been reported in the upper Vallesian at Can Jofresa (Santafe and Casanovas 1978). Thecranial fragmentsofH.siniorrensecome from El Lugarejo, Avila, and belongtoanimmatureindividual; Cerro del Otero. Palencia. classified as Rhinocerossansaiiiensis by Hernandez Pacheco and Dantin (1915); and Toril-3, Daroca-Zaragoza, themostcomplete fragment. As farascan beobserved in these fragments, the nasal notch and theanteriororbital edge reach thesame level asin H. tetradaciyluin (Guerin 1980) and theprocessus postglenoideus and posttympanicus are also in contact. The ratio between head and limbs is similar in .4. siniorren.se and in the extant species Rhinoceros iinicornis. Dicerorliiniis siinialreiisis. and Diceros hiconiis (according to the mean values ofGuerin 1980). Themandible iswideat thesymphysis, even forfemale individuals, as isobserved in the two specimens from Toril-3, Daroca. one with small incisors (IJ (female), and the other one with a much greater I., (male). The upper dentition is larger than in the type specimen from Simone, France, mainly in P'* and Ph There is an increase in size from the upper Aragonian populations to thelower Vallesian, Los Valles de Fuentiduena, Nombrevilla. The same is observed in the postcranial skeleton. The bones are strong and short, mainly the metapodials. but without being massive bones. There are individual morphological variations in the articular facetsand, asin thedentition, sizeisgreater in the lowerVallesian, but thisincrease is not always veryobvious, such as in the astragalus. Bycontrast, there is a great deal ofpostcranial material from the upper Vallesian site ofLa Roma-2, Teruel, whose morphology iscomparable to that ofA. siniorrense but with a significantly different size and robustness (Cerdenoand Alcala 1989). At themoment, .4. alfamhreuse hasonlybeen identified from La Roma-2, butsome ofthe French remains from Montredon, classified ascf. Prosantorliinus (Guerin 1980, 1988), could beascribed to A. alfainhrense. Thisspecies would constitute a final stage in theevolution ofH. siniorreii.se. thespecies from which it would have been derived, from populations like that of Los Valles de Fuentiduena, where the trend ofincrease in size and strength can already be noticed. At the same time, other populations ofA. siniorrense seem to increase in size, but retain their proportions, such as the upper Vallesian samples from Valles-Penedes and the French sites. Hoploaceratherhmi tetraductyhim (Lartet, 1837) H. leiraiiaciyluni has been reported in upper Aragonian sites of the Valles-Penedes basin (Santafe 1978a). Moreover, it is present at three other sites in the Madrid area; Paracuellos-5, where it has been recognized previously by Alberdi et al. (1985), but on the basis ofsome bones that really do not belong to this species. With some doubt, because ofthe scarcity ofremains, //. tetradactyluni is thought also to be present at Cerro dela Plata and Henares-1 (Cerdeno 1989). On theotherhand, itspresencecannot beconfirmed in theVallesian site of Nombrevilla (Santafe et al. 1982) based on the studied sample (mainly dentary remains) which is comparable to that of A. siniorrense from Los Valles de Fuentiduena. PALAEONTOLOGY, VOLUME 300 35 //. tetvadactylum has recently been removed from the genus Aceratherium by Ginsburg and Heissig (1989). One ofthe diagnostic characters indicated by the authors, the semilunate outline ofthe centrale in the tarsus, does not agree with my own observations on the type material of the species from Sansan, France. Aceratherium incisiviim Kaup, 1832-1834 This species is better documented in the Valles-Penedes basin than in the rest ofSpain. Alberdi et al. (1981) recognized Aceratherium cf. incisivum at Los Valles de Fuentiduena. It is also present in the Teruel area, the upper Vallesian sites of La Roma-2 and Masia del Barbo, and the middle Turolian site of Concud. The presence ofA. incisivum at the upper Turolian site ofVenta del Moro, Valencia (Guerin 1980; Morales 1984) cannot be supported. The mandible described by the authors does not show the characteristics of the type material from Eppelsheim, Germany (Kaup 1832-1834). The symphysis is narrower and it is more like Dicerorhinus schleiermacheri from the latter site. So, this implies that the last appearance ofA. incisivum, and thus of the subfamily Aceratheriinae in Spain, corresponds to the middle Turolian. Subfamily rhinocerotinae Owen, 1845 Tribe teleoceratini Hay, 1902 The teleoceratines are not well represented in Spain. There are some reports of the genus Brachypotherium, and the presence of the genus Prosantorhinus has been established for the first time (Cerdeno 1989). Rhinocerotidae indet. II cf. Brachypotherium Roger, 1904 The oldest remains in the Spanish Neogene related to this group are a few bones and a fragment ofan upper F from the Lower Miocene site ofLoranca del Campo, Cuenca. These are scarce elements and not significant enough for an accurate identification. Brachypotherium aurelicmense (Nouel, 1866) Some remains from Rubielosde Mora, Teruel, classified as Brachypotherium sp. (Aguirre and Moissenet 1972) seem to correspond to B. aurelianense, like those of Moli Calopa (Santafe 1978a). Adding to these reports, B. aurelianense has been identified in LaArtesilla, Zaragoza, from a few remains that includea wide and short astragalus whose size and proportions indicate its identity. Rubielos de Mora and Moli Calopa are younger sites than Loranca del Campo within the Lower Miocene, zone A, while La Artesilla is even more recent, corresponding to a lower Aragonian age, the lowest part of zone C, and this indicates the most recent known remains ofB. aurelianense in Spain. Santafe and Belinchon (1988) noted the presence of5. aurelianense at Bunol, Valencia, but these remains are Prosantorhinus sp. (see below). Antunes and Ginsburg (1983) believe that B. aurelianense must be included in the genus Diaceratherium, keeping the genus Brachypotheriumjust for the youngest species B. brachypus. I have in preparation a revision of the French Miocene material of brachypotheres, and I prefer now to maintain B. aurelianense in Brachypotherium. Brachypotherium brachypus (Lartet, 1837) B. brachypuswasreported insome upperAragonian Spanish sitesofthe Daroca area, Zaragoza (Guerin 1980). The fossil material preserved at Utrecht University indicates the presence ofthis species at Arroyo del Val-4, Manchones-1, and Manchones-2, despite the general scarcity of\remains. Prosanthorhimis sp. As noted above, another teleoceratine genus, Prosantorhinus. has been recognized for the first time in Spain. Checking the lower Aragonian material from Bunol, I found a McIV which caught my attention because of 1 CERDENO: SPANISH RHINOCEROSES 30 its great robustness and small size. This bone is not coincident in size or proportions with the French B. aurelianense. However, a comparison with two MclV casts of Prosantorhhms gennanicus from Sandelzhausen, Germany, supports the idea ofthe presence ofProsantorhimis at Bunol. Certain differences do not permit the identiheation ofthe McIV from Bunol with the German species but, on the other hand, there issome unpublished French material identified as Prosantorhimisdoiivillei(Antunes and Ginsburg 1983, p. 24) with which the Bunol material should be compared. The McIV from Bunol also presents similarities with the Lisbon material classified cither as Diuccrathcriwn aiirelianensis or as Guindatheriwn rexmaniudi (Antunes and Ginsburg 1983). The identification ofthe McIV as Prosantorhimis caused doubts about the correct classification of the Portuguese material. Furthermore, Prosantorhimis was also indicated at Lisbon from some dental remains. The Lisbon material must be revised in detail because some ofthe bones classified as D. aiirelianensis could belong to Prosantorhimis. Besides, the presence ofthe genus Gaindatherium in Lisbon seems very doubtful. A brachypodial postcranial skeleton has neverbeen associated with thisAsian genus, contrary towhat isstated for Pro.santorhimis(Heissig 1972, 1974), and possibly the Portuguese postcranial bones assigned to Gaindatherium belong to Prosantorhimis. Present data showthat Prosantorhimiswasdistributed inWestern Europein sitesin Portugal (Lisbon), Spain (Bunol), France (Baigneaux, Artcnay, Bcaugency, Savigne, La Romieu), and Germany (Sandelzhausen, Steiermark, Georgensgmiind),coexistingin somecaseswith B. aurelianense(La Romieu, Savigne). The Middle MN MN Miocene age of these sites ranges from unit 3 (Savigne or Les Beilleaux) to 5/6 (Georgensgmiind). Tribe rhinocerotini Owen, 1845 Subtribe elasmotheriina Bonaparte, 1845 The third group of rhinocerotids recorded in the Spanish Neogene corresponds to the elasmotherines. This group haschanged its rank from family to subtribe, according to Prothero and Schoch (1989). Hispanotheriiim matriteuse (Prado, 1864) The elasmotherines evolved mostly in Asia, but in the Middle Miocene the species Hispanotheriiim matriteuse appears in the Iberian Peninsula and recently it has been found in France (Ginsburg et al. 1987«). First the species (Prado 1864), and later the genus (Crusafont and Villalta 1947) were recognized from dental remains from Puente de Toledo, Madrid. These teeth were characterized by their tendency toward hypsodonty, undulating enamel, and much cement. H. matritense has now been recognized in Spain at Dchesa de los Caballos, Plasencia-Caceres; Torrijos, Toledo; Corcoles, Guadalajara; the Daroca area, Torralba de Ribota, and Tarazona de Aragon, Zaragoza; and recently at La Retama, Cuenca. The postcranial bones are small and slender and they show great similarity with H. griimni, another species described by Heissig (1976) from the Anatolian Peninsula, Turkey, but larger and slightly less slender. Other forms with clear affinities to Hispanotheriiim have been described from the Miocene ofAsia. Antunes andGinsburg(1983) haveconsidered themassynonymsofHispanotheriiim at generic level, while Forteliusand Heissig (1989) place them in Begertheriinn. Thegeographicdistribution ofHispanotheriiim ledAntunes (1979) toproposea migration route through the Alpine Arch from theAsian regions to the Iberian Peninsula. This route hasalso been proposed forthe bovids of the tribe Boselaphini (Moya-Sola and Alferez, in press). Later, Hispanotheriiim may have reached France from Spain. theH.zomnaetrMitNens4e6eoxirstDe;d Cbroirecfolyleisn ctohuelIdbebreiaonldPeerniannsdulas.eeImnsStpoaibne,iitncilsurdeesdtriicntezdonien tChe(lmoiwdedrleArAargaognoinani)a.n Itno MN Portugal, Hispanotheriiim matriten.se appears in younger beds corresponding to the 5 unit, as well as in France. On theother hand, theAnatolian species has a greater time span through thewhole upper Aragonian. Fortelius and Heissig (1989) present a cladistic analysis ofthe elasmotherine group, in which Caementodon is very close to H. matritense, confirmed by my own observations (Cerdefio 1989). I agree also with these authors when they remove Shennotheriiiin hipsodontiis(Huangand Van 1983) from the Elasmotherina because this species does not even show the dental characters ofthe group. PALAEONTOLOGY. VOLUME 302 35 Tribe rhinocerotini Owen, 1845 Subtribe dicerorhinina Ringstrom, 1924 The taxonomic rank of this group has varied (Heissig 1989; Prothero et al. 1986; Prothero and Schoch 1989), even if the subfamily rank is retained by some authors (Guerin 1989). Within Dicerorhinina, theattribution ofspecieslikeL.sansaniense, D. schleiermacheriorS. miguekrusafonti has also been discussed. For the first of these species I use the generic name Lartetotherium as defined by Ginsburg (1974), and I agree with Groves (1983) when he considers that the other European species must be separated from the recent genus Dicerorhinus {D. sumatrensis). Dicerorhinina indet. Theoldest representative ofthedicerorhines in Spain isa rhinoceros from the lowerAragonian ofLaArtesilla, Zaragoza. The remains are scarce and non-diagnostic. They are identified as Dicerorhinina because of their global similarities with Dicerorhimis montesi from Buhol (Santafe et a!. 1987). This rhinoceros from La Artesilla corresponds to the Rhinocerotidae indet. Ill ofText-figure 1. Lartetotherium sansaniense (Lartet, 1851) Several Spanish sites have provided scarce remains ofLartetotherium sansaniense: Paracuellos-3, Madrid; La Cisterniga, Valladolid; Cerro del Otero, Palencia; Coca, Segovia; and Brihuega, Guadalajara, in the upper Aragonian; and Relea, Palencia; Can Ponsic, Barcelona; and perhaps Nombrevilla, Zaragoza, in the lower Vallesian. Study of the whole complex, and comparison with the type material from Sansan, France, led to the realization that the dentition is basically identical, but the postcranial skeleton isdifferent (Cerdeno 1986). All comparable bones from the named sites are smaller than the bones from Sansan, and only a pyramidal and acuboid from Paracuellos-5, Madrid, arecloser to their French homologues. These latter two bones were first classified as Aceratheriwn tetradaetylwn (Alberdi et a!. 1985), but a later revision suggested a Dicerorhinus morphology. It isnoteasy tointerpret these Lartetotherium remains. Sansan, aswellasParacuellos-5,areolder sites than the others and so there could have been a decrease in size with time. Added to this material of L. sansaniense, there are some other older remains (lowerAragonian, zone C) whose morphology and size are also closer to Sansan and their relationship is difficult to establish. This is the case for the remains from Can Mas (Santafe 1978; Cerdeno 1986) and the Dicerorhinus montesi from Bunol (Santafe et al. 1987). Following theseauthors, D. montesiisrelated to D. schleiermacherifrom theUpper Miocenebut itcould bemoredirectly related to Lartetotherium sansaniense, and even in the same genus. New material ofD. montesi (the skull and dentition are still unknown) is needed to support this possible relationship. Dicerorhinus schleiermacheri (Katip, 1832-1834) Anotherclassical speciesofDicerorhinus, D.schleiermacheri,mightalso berelatedtoLartetotherium.This large species is well represented in the Upper Miocene of Western Europe. In Spain, it is known in the lower Vallesian ofthe Valles-Penedes basin and from several localities in the upper Vallesian ofthe Teruel area like Masia del Barbo and La Roma-2. The abundant bones from La Roma-2 are very large and some of them surpass the maximum values established by Guerin (1980) for this species. Their robustness is also somewhat greaterthan atothersites. IntheSpanish lowerTurolian, D.schleiermacheriispresentat Piera, Valles-Penedes; Puente Minero. Teruel; and Crevillente-2, Alicante. At this last site, the species is very well represented by dental remains, but only three bones have been recovered, contrary to what occurs at La Roma-2 wherejust six teeth have been found. The species is present through the rest of the Turolian, but is much scarcer. It is known from the middleTurolian ofConcud, Teruel, and the upperTurolian ofLas Casiones and El Arquillo, Teruel; Venta del Moro, Valencia; La Alberca. Murcia; and Los Hornillos and El Fargue, Granada. As already commented, the classification of the mandible from Venta del Moro as Aceratherium incisivum (Morales 1984) has been modified (Cerdeno 1989). Together with the mandible, there is a very large cuboid that can be assigned to Dicerorhinus schleiermacheri. Guerin (1980) classified the metapodials from El Fargue as Dicerospachygnathus (he identifies this species at Cenes dc la Vega, but the described material comes from El Fargue). This taxonomic determination is not a . 1 CERDENO: SPANISH RHINOCEROSES 303 AGE STAGE 1 2 LOCALITY 1 a b c d e f g h j k 1 m n 0 PLIOCENE RUSCINIAN 15 Layna k 14 La Catera k Et Fargue k Cenes de la V. k UPPER 13 La Alberca k TUROLIAN Venta del More k Las Casiones k El Arqui1lo k MIDDLE Crevi1lente 15 k 12 Los Mansuetos * k TUROLIAN Concud k UPPER LOWER Piera * k MIOCENE 11 Puente Minero k TUROLIAN Crevi1lente 2 k Can Perellada * UPPER 10 Can Jofresa * VALLESIAN Masfa del Barbo * k La Roma 2 k k I Can Llobateres k 7 LOWER Can Ponsic * -* 9 Nombrevi1la * 7 VALLESIAN Chiloeches * H Los Valles de F cf Relea * Brihuega * Coca * 7/8 La Cisterniga * UPPER G Cerro del Otero * Andurriales * Tori1 3 * 6 Arroyo del Val ARAGONIAN Manchones 1 y 2 k Armantes 3 k MIDDLE F Paracuellos 3 k * Paracuellos 5 cf k E 5 Henares 1 7 Puente de Toledo MIOCENE MIDDLE Tarazona Torrijos k ARAGONIAN D 4b Munebrega 1 k Valdemoros 1A k Torralba V k Corcoles k LOWER Bunol li^ 7 C 4a Can Juli ARAGONIAN Can Mas 7 B Artesi1la 11 4r A Rubielos de Mora LOWER 3 Molf Calopa RAMBLIAN Z Loranca 11=cf •k MIOCENE Y 2b Valquemado * 2a Cetina de A. I TEXT-FIG. 1. Biostratigraphical distribution ofthe Rhinocerotidae in the Spanish Neogene localities. I, Daams and Freudenthal (1981). 2, Mein (1975). /, Rhinocerotidae indet.; a, Brachypotherium aiirelianense/B. brachypi(s\ h, Prosantorhiiwssp.; c, Plesiacercitheriumplatyodon: cl, P. mirallexi: e, Protaceratherium minutunv, f, Alicornops simoirense', g, A. alfamhrcii.se', h, floploacerallieriiim Ictnulactylum,j, Acerallieriiim incLsiviim: k, ffi.spano!lierium matriteusc, I, Lartctothcriwn .saiLsanicii.se', m, Diccrorliiiiu.s iiiontesi', n, DiccrorliiniLS sclileienmicheri a, Slcphanorliinits migncicni.safonti. ', PALAEONTOLOGY, VOLUME 304 35 Justified. D. pavhygnathus has been defined on material from Pikermi, Greece. Bones from this site (Musee Guimet d’Histoire Natiirelle. Lyon) show that differences from D. schleiennacheri from La Roma are limited to the greater robustness of the Greek material. On the other hand, Geraads (1988) has revised the rhinocerotids from Pikermi, and has noted the presence of two forms; Ceratotherium neiimayri (= Diceros pachygmithus) and Dicerorhimts pikerniiensis (= Dicerorhiniis schleiermacheri var. orientalis). This author established cranial differences between these two species, but he could not do the same with the postcranial skeleton (two different morphologies are observed only for some bones). The similarities between the metapodials from El Fargue and Diceros pachygnathus are logical since those of D. pachygnathus could correspond to the Dicerorhiniis ofPikermi, a form very close to Dicerorhiniis schleierniacheri, but less slender. Therefore, the bones from El Fargue must be classified as Dicerorhiniis cf. schleiermacheri (Cerdeno 1989, p. 350). The main difference between thesemetapodialsand theothersisthegreaterrelativelengthofthe Mtlll. Stepluinorhiinis migiielcrusafonti Guerin and Santafe, 1978 5. inigiielcriisiijonti is the last species studied among the Spanish Neogene dicerorhines. It has also been described as Dicerorhiniis, but it corresponds to the Plio-Pleistocene group named Stephanorhinus (Groves 1983). This species was defined from the Ruscinian Spanish site of Layna, Soria (Guerin and Santafe 1978), and also recognized at Perpignan, France. Furthermore, the species isidentified at La Calera, Teruel (Cerdeno 1989). The skull and mandible are still unknown. The anatomical comparative study of S. migiielcrusafonti shows the different slenderness ofthe Spanish metapodials compared to the Perpignan ones, which have the highest indices. Besides this, there are few morphological differences between S. megarhinus and the Plio- Pleistocene species. The Spanish rhinoceros can be distinguished from thecommon Pliocene European species S. megarhinus, mainly by its smaller size. The robustness is comparable or slightly less than in S. megarhinus, and only the Perpignan bones are stronger than all other Plio-Pleistocene species. The best knowledge ofS. migiielcrusafonti could lead to two contrary results. Either it is fully confirmed as a different species, or it becomes a variety ofS. megarhinus, possibly a subspecies. At the end ofthe Ruscinian, both forms disappear. CONCEUSIONS Nine genera and fourteen species of rhinoceros have been established in the Spanish Neogene. Protaceralheriiim mimitiim and Prosantorhiinis sp. are reported in Spain for the first time, and AUcornops alfambrense appears in the upper Vallesian as a final stage of the A. simorrense evolutionary lineage. The Spanish Miocene and Lower Pliocene (Ruscinian) are quite well documented for rhinoceroses, and only two biozones are poorly known for macromammals (not only for rhinoceroses); zone MN 2a or X in the Lower Miocene and zone MN 3/4 or B in the early Middle Miocene (lower Aragonian). Spanish rhinocerotids document several stages that can be grouped as follows: MN 1. Lower Miocene, zones 2/>-MN 2a or Y-Z. Characterized by the abundant presence of Protaceratheriiim mimitum. This species is coeval with cf. Brachypotherhim at Loranca del Campo. MN 2. Lower-Middle Miocene, zones 3/?-MN 4aor A-(B)-C. Corresponds to a period ofhighest diversity, but with a rather low numerical representation. There are six species, although no more than four at the same site (e.g. Bunol). MN 3. Middle Aragonian, zone 4bor D. Hispanotherhim matritense is the only rhinoceros known in the sites of this age, very abundant in localities like Corcoles or Torrijos. MN 4. Middle-upper Aragonian, zones 5/6 or E-F. Rather poorly documented, and apparently only recognized in the Madrid area. Rich sites for macromammal fauna such as Paracuellos-5 have provided only a very few remains of rhinocerotids, compared to the abundance of the equid Ancliitlieriiim. 5. Upper Aragonian-Vallesian, zones MN 6-MN 10. AUcornops simorrense is characteristic at this time. It is widely distributed in Spanish basins with its best representation in Western Europe. It can coexist with Brachypotherium brachypus or Lartetotheriwn sansaniense, but always as the predominant species. A. simorrense does not appear in Spain during the upper Vallesian, except in the one site ofCan Jofresa, Valles-Penedes basin. However, a probable descendant, A. alfambrense. CERDENO; SPANISH RHINOCEROSES 305 QUATERNARY u Hunnldi-Tenipe- ty rature dry wetworrcold 7 I !./m. 16 Villafranchian S E 1 II45 Ruscinian g g o 1 tOP; co 13 12 Turolian I# I 1 c 910 VaIIesian O W> JD O 76/8Fr® upper X [ 5 E middle ^Q.E o o D Q_'CL'(tQQ I C < B A Ramblian I <) Z Icf. YX Agenian r > OLI G 0 C ENE TEXT-FIG. 2. Schematic representation ofthe global diversity ofthe Spanish Rhinocerotidae compared with the diversity ofEquidae and global palaeoclimatic changes. is present at La Roma-2, Teruel basin. In the rest of Western Europe, A. simotrense has been identified in the upper Vallesian, but it is possible that a part of this material, as well as cf. Prosantorhinus sp. D from Montredon, France (Guerin 1980, 1988), could be related to A. alfanihrense. 6. Turolian-Ruscinian, zones MN 1 1-MN 15. Characterized by the decline of the family Rhinocerotidae. Rhinoceroses are abundant in macromammal faunas until the upperVallesian, but after the Turolian they become more and more scarce. Despite this, rhinocerotids persist until the Upper Pleistocene. The passage between the last two stages is gradual. In fact, the most characteristic Turolian form, Dicerorhimis schleiermacheri, is already present in the upper Vallesian, together with A. incisiviim. This latter species is occasionally found in the lower and middle Turolian. As early as the latest Miocene (MN 13), D. schleiermacheri is the only extant rhinoceros which is replaced by Stephanorhinus miguelcriisafouti in the Ruscinian. This one, in its turn, is the only representative of the family at that time and is replaced by S. etruscus in the Villafranchian. All these different stages give us a global temporal distribution which is compared with that of the equids (Text-figure 2). The greatest diversity of the Rhinocerotidae coincides with the development ofjust one equid, Anchitherium aurelianense, a species that hardly varies through the Aragonian. The arrival of the equid Hipparioii, among other immigrants, occurs at the same time that rhinocerotids begin to decline. The global distribution of the rhinocerotids can be correlated with the palaeoclimatic curves established by Lopez et al. (1987) for the Spanish Neogene (Text-fig. 2), mainly based on micromammals. The predominance of Protaceratherium miinitiim coincides with a warm and wet period which becomes gradually drier and colder. Daams and Meulen (1984) have also established from micromammal faunas a humid environment during the transition Agenian-lower Aragonian. The next stage, with higher diversity, corresponds to both cold and wet maxima in the palaeoclimatic curves. Hispanotherium matritense occurs when the conditions become drier, the temperature increases, and the humidity reaches a minimum. This species was well adapted to arid conditions, and it seems that the environment was more advantageous for this rhinoceros than for . 306 PALAEONTOLOGY. VOLUME 35 Othermacromammals like the equid Anchitheriiim. This would explain their different representation in the middle Aragonian sites. Later, palaeoclimatic curves show that Alicomops simorrense apparently lived in a moderate climate with a global tendency to cold weather with humidity oscillations. The last period ischaracterized once again by maximum aridconditions in theTurolian which decrease towards the Ruscinian. Possibly, the Turolian aridity favoured Hipparkm over rhinoceroses which do not regain their previous diversity. Ack)unvledgements This work has been mainly supported by a grant from the CSIC within the Research Project El Plio-Pleistoceno de la cuenca de Guadix-Baza y el corredor Huercal-Overa: evolucion faum'stica y geodinamica. 1 thank several colleagues for their critical reading of the manuscript. I also thank Dr Kurt Hcissig (Universitiit Institut fur Palaontologie und historische Geologic. Mtinchen) for the useful casts he provided. REFERENCES AGUIRRE, E. and MOissENET, E. 1972. Precisions sur le gisement Miocene de Rubielos de Mora (Teruel, Espana). Melanges de la Casa de Veidzejuez, 7, 561 -563. ALBERDi. M. T., CERDENo, E. and HERRAEZ, E. 1985. Pcrissodactyla de la provincia de Madrid. 61-80. In Geo/ogi'a y Paleontologi'a del Tereiario continental de laprovincia de Madrid. Servicio de Publicaciones del C.S.EC., Madrid, 105 pp. GiNSBURG, L. and MORALES, J. 1981. Rhinocciotidae del yacimiento de los Valles de Fuentidueha (Segovia). Estudios Geoldgicos, 37, 439-465. ANTUNES, M. T. 1979. Hispatiotherium fauna in Iberian Middle Miocene; its importance and paleogeographical meaning. Annales Geologiqiies du Pays Hellenicpie. Vlltli International Congress on Mediterranean Neogene, Athens, 19-26. and GINSBURG, l. 1983. Fes rhinocerotides du Miocene de Lisbonne-systematique, ecologie, paleobiogeographie, valeur stratigraphique. Ciencias da Terra (Universidade Nova de Lisboa), 1, 17-98. BELiNCHON, M. 1987. Estudio tafonomico y sistematico de la fauna de macromamiferos del Mioceno de Bunol (Valencia). Unpublished Ph.D. thesis, Universidad de Valencia. BONIS, L. DE 1973. Contribution a I’etude des mammiferes de I'Aquitanien de I’Agenais. Rongeurs, carnivores, perissodactyles. Memoires da Museum National d'Histoire Naturelle de Paris, 28, 1-192. CERDENO. E. 1986. El csqucleto postcraneal de Lartetotlieriuin sansaniensis (Mammalia, Rhinocerotidae). Estudios Geoldgicos, 42, 197-209. 1989. Revision de la sistenidtica de los rinocerontes del Nedgeno de Espana. Editorial de la Universidad Complutense de Madrid, 429 pp. and ALCALA, L. 1989. Aceratherium alfambrense sp. n., nueva especie de rinocerotido del Vallesiense superior de Teruel (Espana). Revista Espatiola de Paleontologi'a, 4, 39-51. CRUSAFONT, M. and viLLALTA, J. F. 1947. Sobrc un interesante rinoceronte {Hispanotherium) del Mioceno del Valle del Manzanares. Las Ciencias, 12. 869-883. CUVIER. G. 1822-1825. Recherches sur les ossementsfossiles. Dufour et D’Oeagne, Paris. DAAMS, R. and FREUDENTHAL, M. 1981. Atagonian the stage concept versus Neogene mammal zones. Scripta : Geologica, 62, 1-17. and VAN DER MEULEN, A. 1984. Palaeoenvironmental and palaeoclimatic interpretation ofmicromammal faunal succession in the Upper Oligocene and Miocene ofNorth Central Spain. Interim colloquium of the R.C.M.N.S. on paleoclimatic evolution, 241-257. DOLLO, L. 1885. Rhinoceros vivants et fossiles. Revue des Questions Scientifiques, 17, 293-299. FORTELius, M. and HEissiG, K. 1989. The phylogenetic relationships of the Elasmotherini (Rhinocerotidae, Mamm.). Mitteilungen Bayerische Staatssammlungen fiir Palaontologie und Historische Geologie, 29, 227-233. GERAADS. D. 1988. Revision des Rhinocerotinae (Mammalia) du Turolien de Pikermi. Comparaison avec les formes voisines. Annales de Paleontologie, 74 (1), 13^1. GINSBURG, L. 1974. Lcs rhitiocciotides du Miocene de Sansan. Comptes Remlus de TAcademic des Sciences de Paris, 278, 597-600. and GUERIN, c. 1979. Sur I’origine et fextension stratigraphique du petit rhinocerotide miocene